Mobirise

OCARINA - Optimal Control frAmework for Robust Iterative NonlineAr experimental design

ANR project 25-CE48-2184

About

Experimental design focuses on optimizing data acquisition and use so as to limit the experimental burden. For parameter identification in model-based input–output systems, a classical objective is to reduce estimation uncertainty, a goal that can be formulated within optimal control. Project OCARINA investigates how control-theoretic tools can advance experimental design.
To incorporate prior information, experiments are performed in closed loop, with each iteration building on the current estimate. Modeling the unknown parameters as distributed variables enables the explicit treatment of uncertainty through ensemble control. Leveraging optimality principles in this setting leads to new strategies where uncertainty is directly embedded in the construction of optimal experiments.
The project is motivated by a neuroscience application involving retinal stimulation and recording. In this context, determining appropriate inputs requires a calibration stage, while the short lifespan of the preparations makes such tuning particularly costly. A central objective of OCARINA is to deploy the proposed methods in this setting, both to alleviate these practical limitations and to clarify the broader issues underlying experimental design.

People

Consortium.
Collaborators.

Publications

‌

Best AI Website Maker